131–140 of 209 posts
Earlier quoted context omitted.
There is some very deep knowledge in SO's comments and some lower rated replies. But I suspect that is not what was making Stackoverflow so popular.
I suspect the new Stack Overflow is Discord channels, at least for some platforms/frameworks. Too bad the free-form non-threaded discussion format of Discord makes things so hard to follow.
We were (finally!) given the go for even using AI just before summer vacation this year, and I was very excited, having been obsessed with AI 20 years ago. I was still excited quite a while, until I slowly understood all the limitations that come with LLMs. We can not trust them, and this is one of the fundamental problems: verifying things takes a long time, sometimes its even faster just writing the code yourself.
We do have non-core-product tasks that can greatly benefit from AI, but that is already a small part of our job.
I did find two areas where LLMs are very useful: generatin documentation from code (mermaidjs is useful here) and parsing GDB output!
Seriously, parsing GDB output was like an epiphany. I was for real blown away by it. It correctly generated a very detailed explanation of what kind of overwriting was happening when I happened to use a wild pointer. Its so good at seeing patterns, even combining data from several overwritten variables and parsing what was written there. I could have done it myself, but I seldom do such deep analysis in GDB and it did it in literally 10 seconds. Sadly, it was not that terribly useful this time, but I do feel that in the future GDB+AI is a winning concept. But at the same time, I spend very little time in GDB per year.
Morale is low because leaders think AI can do that amount of work, but it can’t actually (at least not yet). This both means that they don’t hire enough people to do the work needed, while also “drive by” insulting the intelligence of the people they are overworking.
For a good example, just look at how Google does "support." It's just robots doing shoddy work and screwing over people. Can better compensated and organized human support team do better? Of course, but the rich execs don't want to spend a penny to help people if they can get away with it.
Earlier quoted context omitted.
How can AI search into documentation, if the documentation is a thousands of obsolete and contradicting Jira tickets, few outdated Confluence pages with mail attachments and handful of excel files on SharePoint?
With the necessary MCP servers.
Earlier quoted context omitted.
Are you saying everyone who isn't barely starting their career is a genius? In the current state of things I'd gladly take mediocre work from a human over slop from an AI.
Seriously this. Doing code reviews on LLM created code is so frustrating. If the code was submitted by a junior engineer I could get on zoom with them and educate them, which would make them a better team mate, advance their career goals, and make the world slightly better. With AI created code the review process is a series of tiny struggles to dig up out of the hole the LLM created and get back to baseline code qua…
Earlier quoted context omitted.
You don't use any AI - drafting documentation, write boilerplate code, transcribing meetings, simplifying team communications, searching product documentation, alternative to Google or StackOverflow, creating presentations, as a brainstorming partner? I would consider all of these "work". If you say AI does 0% of your work, I'd say you're either a genius, behind the curve or being disingenuous.
How can AI search into documentation, if the documentation is a thousands of obsolete and contradicting Jira tickets, few outdated Confluence pages with mail attachments and handful of excel files on SharePoint?
Earlier quoted context omitted.
With the necessary MCP servers.
Good luck to receive security clearance for that. Even Cursor is not allowed, though everyone is using it with private account.
And that's coming from someone that has repeatedly gone on record saying "my expectation for our industry is a gigantic contraction because of LLM", ...but this isn't a scenario that's plausible with current models.
AI has drastically changed how I make decisions about code and how I code in general. I get less bogged down with boilerplate code and issues, which makes me more efficient and allows me to enjoy architecting more. Additionally, I have found it extremely helpful in writing lower-level code from scratch rather than relying on plug-and-play libraries with questionable support. For example, why use a SQLite abstraction…
Because of the accumulated knowledge in these abstraction layers and because of the abstraction itself resulting in readable and maintainable code.
Yes you can move the abstraction one level up, but you don't control it if you nor the LLM meet the level of accumulated knowledge that is embedded in this abstraction. Let alone future contributors to your codebase.
Of course it is all depending on context and there is no one-fits-all strategy here.